With the close of early voting, we say the traditional bump on the final days with over 4000 cast on Friday in the Democratic primary. A total of 15,826 votes were cast early (with a slight bump in that eventually to come as some mail ballots are received before the deadline).
As such, here's what the total Travis county Democratic primary turnout will look like depending on the percentage of the total vote comes from early vote compared to election days. Here is what total turnout looks like in a couple different scenarios. The bolded one is what our BOR models had been based on though I've provided a number of different scenarios.
If 45% of the vote is cast early... Dem Primary turnout is: 35,169
If 48% of the vote is cast early... Dem Primary turnout is: 32,971
If 52% of the vote is cast early... Dem Primary turnout is: 31,652
If 55% of the vote is cast early... Dem Primary turnout is: 28,775
If 60% of the vote is cast early... Dem Primary turnout is: 26,377
This is the final chart of how those models turned out. The final day's vote of over 4,000 was a little higher than expected. For history's sake...
For the first time in years, we had a weather related event that had an appreciable and noticeable affect on early voting turnout. It's so much so, that I've had to adjust the formulas for the final days to discount the 60-70% dropoff that occurred on Tuesday due to the "snowstorm" that blanketed the city in a quarter inch of wintery mix. Still, polls were closed for 4 hours and opened later the next day.
All that being taken into account, there are probably about 1000 Democratic primary votes that disappeared from the universe as a result after you factor in the folks who will shift their voting to another day. The new Democratic primary turnout models for Travis County are below.
Today's Estimate: 31,424 Total Democratic Travis County Primary Voters
(-2275 from Monday's projection)
You know, I'm starting to think that at the end of all this, it's possible that the combined model for all past elections could be the most accurate one in addition to the fact that it's been stable almost the entire time.
Not a lot of movement in the top voting locations. The Tax Office managed to bump up and take over the #2 slot from the Randallas Ben White which is now #3. Randalls Research still leads with 791 total votes cast.
The bottom 5 turnout locations are the same 5 locations but have reordered. Thanks to Bill White's rally Wednesday night at Parque Zaragoza, the daily turnout jumped from about 20 a day to 86 which was enough to push it up one spot in the ranking past the MT Chinatown Supermarket. The Dan Ruiz Library improved one position and swapped places with the Carver Museum which is now tied for worst turnout with the Goodwill at Norwood.
The UT campus site remains steady at #9 in the middle of the pack.
Fun note: While the low turnout Democratic polling sites may have only seen 136 voters, the Republicans in Travis County have 4 polling locations with even lower turnout (even though they have higher overall turnout). For the GOP, the Carver Museum has maintained it's crown as lowest early vote turnout polling place with just 16 votes, followed by Parque Zaragoza with 21. And no, there was no GOP turnout bump at Parque Zaragoza as a result of Bill White's rally there.
It's been a few days since I've pulled together the early vote numbers and run them through the models for projecting total Democratic turnout for the primary (at least in Travis County). The projections had been falling every day so far until today, when it actually bounced up due to a higher than expected Monday turnout. So it's possible that we may have hit a stabilizing point or it could just be that people have enjoyed the nice weather and are more in a mood to vote. With some degree of wintery mix hitting the Central Texas region Tuesday, it will be interesting to see if that dampens the recent uptick in voters.
Today's Estimate: 33,699 Total Democratic Travis County Primary Voters
(-2331 from Friday's projection, +602 from yesterday's)
Remember, follow the red line for our purposes here, it's the weighted average primary model and it's nice to see everything else slowly closing in on it. According to the models, we should see about 1450-1500 Democratic primary votes cast in Travis County for Tuesday, but that's assuming that the cold snap doesn't have an impact which I assume it will.
Top early votes for Democrats by turnout are Randalls Research (620), Randalls Ben White (547), Tax Office (543), Randalls South Mopac (518), & Ben Hur Hall (480). The UT location which started out first and then fell to fifth in our last report, has kept dropping and is now 9th overall, though that is in part to traditionally weak weekend voting on campus and traditional strong weekend voting at places people shop.
The overall weakest early vote sites for Democrats are Goodwill Norwood Park (108), Dan Ruiz Library (110), Carver Museum (111), Parque Zaragoza (116), and Chinatown Center (119).
In somewhat exciting news, Republican poll workers at Parque Zaragoza finally saw a primary voter for the first time Monday since last Thursday bringing the total GOP early vote at that location to 14. But don't be fooled, the lowest GOP early vote site turnout crown still belongs to the Carver Museum which has seen just 12 GOP voters grace their doors.
Yesterday's voting came in lower than projected which in turn will lower today's projected estimate for the Democratic primary turnout in Travis County. Additionally, Travis County revised their numbers from the first two days downward which is reflected in today's estimate but not corrected for the first two so it makes the drop appear bigger.
Today's Estimate: 36,030 Total Democratic Travis County Primary Voters
(-1544 from yesterday's projection)
Daily Projection by Model for Total 2010 Travis County Democratic Primary Turnout
Early Vote is running a little over twice what it was in the last gubernatorial election in 2006 for Democrats. The Republican turnout in Travis County is running higher than the Democratic turnout and is about 3 times what the GOP turnout was in 2006.
The University of Texas had the highest Democratic primary turnout of all voting locations on the first day of early vote and is the only place that has over 100 Dem primary votes cast at a location in any of the first 3 days. It has since dropped to the 5th highest turnout location.
The highest Dem turnout locations are the Tax Office (269), Randalls Research (230), and Ben Hur Shriners Hall (210), Randalls Ben White (208), and UT (205). The lowest Dem turnout locations are the Goodwill at Northwood Park Blvd (45), Parque Zaragoza (54), Carver Museum (58), & Dan Ruiz Library (63).
Winner of the loneliest voter award? GOP primary voters at the Carver Museum which had 1 voter Tuesday, 1 voter yesterday and 4 voters today.
In the past couple of elections, I have posted data on what the turnout is estimated to be in Travis County based upon past trends during early vote. After a couple days of data, correcting some errors, and looking at some historical data for the early vote to election day split I've improved the models and am ready to publish my first estimate of Travis County Democratic Primary turnout.
Today's Estimate: 37,574 Total Democratic Travis County Primary Voters
(+558 from yesterday's projection)
Daily Projection by Model for Total 2010 Travis County Democratic Primary Turnout
With regard to the graph, the blue lines are all past Democratic Primaries (04, 06, & 08) and collectively drive the Combined Primary Model in red which is the best indicator for our purposes this election cycle. Past General Elections are in green and you can see how the heavy turnout years of 04 and 08 result in outliers with low turnout projections.
This is all just a factor of how voters behave in when they vote in the early vote period, which varies in primary and general elections. In the past 3 primaries, the very last day of early voting regularly has twice as many votes cast on that day than on the second to last day. In general elections, while there is still a spike on the last day it's always less than a 50% increase over the second to last day. Constitutional Amendment elections indicated in yellow behave much more like low turnout primaries.
In the past I have only projected the total early vote. This year I did a little research and feel comfortable changing up the charts to go ahead and project total primary turnout. The relative share of votes cast early compared to Election Day has always been the most unpredictable measure.
These models are based off of a projected 48% of the total Democratic Primary vote being cast in Early Vote, and 52% cast on Election Day. Here's a chart of the early vote share as a percentage of all votes cast for all Democratic primaries, general elections, and constitutional amendment elections from 2002 - 2009. I placed a best fit line across all these elections for what it is worth and it happens to match up with my gut feeling.
Monday saw 5488 voters come to the polls for all jurisdictions Travis County which was much higher than expected. We're talking by like over 1000 voters!
This has an interesting effect of making me really just go back through the models to see what what that says. Traditionally, the last day of early voting is almost exactly twice that as the second to last day. If that trend is the more powerful one, then the final early turnout will be 2000-3000 votes higher than either of the models would suggest even with the Monday data plugged in. I feel this is probably the case and as such have overridden those number.
In addition, thanks to another local consultant, I've got some real numbers on the percentage of actual City of Austin votes included in amount that is reported by the Travis County election division each day (which include non-city of austin voters from other cities and jurisdictions holding elections in the county). Unsurprisingly, the Austin vote is 83% of that number, which is within a percent of the ration from the 2006 mayor's race. I'm glad to see that some numbers are staying predictable so I'm plugging that into the models now as well.
Expected Total Early Vote for City of Austin by Latest Model Run (Monday)
Based on an predicted 48%/52% split of early vote to election day vote produces the following projected TOTAL votes by model as of Sunday data. This has the greatest potential for error as it's the least predictable part of the model.
2006 mayoral model: 57,696 total votes
2008 council model: 56,092 total votes
Combined Avg model: 56,617 total votes expected for City of Austin
This is the effect that the last two days of early voting can have on overall numbers. Maybe undecideds have decided and are now voting and had been holding off. Maybe weekend field efforts had an effect. Maybe all those weekend McCracken "violation fliers" got some people to the polls. Who knows, but with more data I think that it is now clear that we are going to pass 25,000 early votes within the City of Austin (which we won't be able to absolutely confirm until the Saturday vote or if we're able to get our local consultants to match the data before then.
Sunday saw 1373 voters come to the polls for all jurisdictions Travis County which was higher than expected for Sunday even as the raw vote amount remains unimpressive. That lifted estimates slightly for all models but because there are so few days of voting left, continued to narrow them around the 23,000 total early vote mark.
Expected Total Early Vote for City of Austin by Latest Model Run (Sunday)
Based on an predicted 48%/52% split of early vote to election day vote produces the following projected TOTAL votes by model as of Sunday data. This has the greatest potential for error as it's the least predictable part of the model.
2006 mayoral model: 51,275 total votes
2008 council model: 44,796 total votes
Combined Avg model: 48,281 total votes expected for City of Austin
Saturday saw 2453 voters come to the polls for all jurisdictions Travis County (which is only 26 off from what the combined model predicted for today). That resulted in the 3 models converging a bit towards the average. It is remarkable stable and we're pretty much guarunteed at this point to see total early vote (for just Austin) to be between 20k-25k voters.
Expected Total Early Vote for City of Austin by Latest Model Run (Saturday)
The easy math is to just go with a 50/50 split but again, I'll be conservative and use the 48/52 early to e-day split, and produce the following projected TOTAL votes by model as of Saturday's data.
2006 mayoral model: 50,858 total votes
2008 council model: 43,475 total votes
Combined Avg model: 47,406 total votes expected for City of Austin
Resident commenter and UT math professor Lorenzo Sadun feels that we'll actually shift to more early vote than election day turnout this year, which would result in even lower total projected turnout than listed below. I'm not convinced that will happen as the city electorate has lagged the partisan elections in shifting their vote early. In 2006 the city was 33% early vote and 2008 had only moved to 43% early. Then again, it was about 75% early for last fall's presidential election so it's entirely possible that it will speed up the trend in which case, a 53%/47% early to election day split would result in something closer to just 43,000 votes total.
KXAN does a quick piece on the elections and turnout and opens it with a graphic referencing our Burnt Orang Report projections! Given the stability of our models, I certainly stand by the numbers! Watch it below.
Friday saw 2751 voters come to the polls for all jurisdictions Travis County (which is only 20 off from what the 2006 mayoral model predicted for today). There were also a few corrections to previous days, nothing major, just a few dozen votes updated at random early vote sites.
Expected Total Early Vote for City of Austin by Latest Model Run (Friday)
The easy math is to just go with a 50/50 split but again, I'll be conservative and use the 48/52 early to e-day split, and produce the following projected TOTAL votes by model as of Friday's data.
2006 mayoral model: 51,392 total votes
2008 council model: 43,006 total votes
Combined Avg model: 47,427 total votes expected for City of Austin
Thursday saw 2588 voters come to the polls for all jurisdictions Travis County. That is only 8 more votes than yesterday. That's actually less than predicted by the 2008 model but still more than predicted by the 2006 model. End result- it pushes the turnout projection to a tighter more stable estimate closer to 23,000.
Expected Total Early Vote for City of Austin by Latest Model Run (Thursday)
The easy math is to just go with a 50/50 split but again, I'll be conservative and use the 48/52 early to e-day split, and produce the following projected TOTAL votes by model as of Thursday's data.
2006 mayoral model: 51,369 total votes
2008 council model: 43,644 total votes
Combined Avg model: 47,746 total votes
